Globally Africa currently has one of the highest levels of deforestation, which affects both habitat quality and species diversity (Bodart et al. 2013; FAO 2010; Lawes, Macfarlane & Eeley 2004; Wade et al. 2003). In South Africa the forest biome is the smallest of the eight biomes (Mucina & Geldenhuys 2006). Eastern mistbelt forests, which form part of the southern mistbelt forest group, are naturally fragmented forests with grassland which occur from the Eastern Cape to KwaZulu-Natal (KZN) (Mucina & Geldenhuys 2006). These forests are dominated by Afrocarpus/Podocarpus species (yellowwoods) (Adie, Rushworth & Lawes 2013; Moll 1972). Drakensberg foothill moist grassland and Midlands mistbelt grassland comprise the largest part of the untransformed grassland habitat in the study area and are home to a wide variety of plants (Mucina & Geldenhuys 2006).

The first forest legislation in KZN was passed in 1853, a forest service was started in 1891, and a functional forestry department in 1902 (Fourcade 1889; King 1941). However, most of the forests had already been exploited by European settlers, including colonial woodcutters (Cawe & McKenzie 1989; Cooper 1985; King 1941). Fourcade (circa 1889) was seconded to the Natal Government to prepare a report on the condition of indigenous forests on Crown and Trust lands (Fourcade 1889). His report subsequently led to the decision in 1891 to protect the Crown Forests (Fourcade 1889; Storrar 1990).

Remnants of sawpits can still be seen in Ngeli Forest and others, and although the forests appear to have recovered substantially since protection was afforded (pers. obs.), regeneration is slow and the canopies are generally irregular and areas reduced (Geldenhuys 2007; Lawes et al. 2004; Moll 1972). Following the colonial logging exploitation in the mid-1800s and intermittently to the 1940s, these forests in KZN have been allowed to recover by natural succession processes (Adie et al. 2013). However, as most of these forests are in close proximity to communal lands, subsistence wood harvesting and use of the forest and grasslands to graze cattle has continued (Adie et al. 2013; Lawes, Griffiths & Boudreau 2007; Robertson & Lawes 2005; Wirminghaus et al. 1999; pers. obs.). Despite being protected, generally conservation of these forests and their associated grasslands are difficult (Adie et al. 2013; Cawe & McKenzie 1989; Downs 2005; Wirminghaus et al. 1999).

It is clearly difficult to determine whether the present plant species diversity in Ngeli Forest and nearby forest patches and associated grasslands is similar to what existed before the exploitation that began in the mid to late 19th century, although some of the specimens collected and included in this species list date from that period. Therefore we documented a checklist of the plants of the forests and grasslands in the Weza district (3029DA WEZA), southern KZN, including Ngeli Forest and nearby indigenous forest patches to highlight the biodiversity importance of this area and need for conservation. The list could serve as an important conservation management tool in future. We also reviewed the plants’ status in the Red Data List.

Of the 1554 records included in this summary of plant species for the Weza district, there were 6 lichens (0.4%), 46 bryophytes (3.0%), 58 pteridophytes (3.7%), 6 gymnosperms (0.4%), and 1438 angiosperms (92.5%) (Table 2). The number of plant species was highest in the grasslands, and most of these were herbaceous species (Table 3). Of the angiosperms, 27.3% were monocotyledons and 72.7% were dicotyledons. The ratio of monocotyledons to dicotyledons for Ngeli was 1:2.7. This was higher than reported for Mahwaqa Mountain (1:2.1) (Meter et al. 2002).

Recently, the flora of the Great Winterberg-Amatholes area (7382 km² in extent) was described (Clark et al. 2014). Whilst this region has similar numbers of families for the pteridophytes, gymnosperms, monocotyledons and dicotyledons, the numbers of genera and species are higher, possibly because of the larger area studied compared with the current study area. The number of endemics recorded in this study (43) exceeds the number of endemics in the Clark et al. (2014) study (35) in a much larger area.

Only plant families with more than 1% of the total number of species for the Ngeli quarter-degree square were included (Table 4). To make this comparable with a similar list (Meter et al. 2002), the family structure of Asclepidaceae before consolidation into Apocynaceae was used (Table 4). The 24 families contributed 71.0% of the species and 57.2% of the genera of this list.

As is the case with Mahwaqa Mountain, KZN (Meter et al. 2002), the most species-rich family for 3029DA WEZA was Asteraceae, with 239 species (Table 4, Online Appendix 1). There were 14 endemics (5.9%) and 67 near endemics (28.0%) amongst the listed Asteraceae. Amongst the Asteraceae, the following species fell into ‘endangered’ categories in the RDL: Berkheya pannosa (rare, endemic); Felicia wrightii (critically rare, endemic); Gnaphalium griquense (rare, near endemic); Helichrysum tenax var. palidum (rare, near endemic); Phymaspermum villosum (rare); and Senecio poseidonis (threatened, near endemic) (Table 5). Asteraceae also included 11 (4.6%) naturalised species.

The second most species-rich family for 3029DA WEZA was Fabaceae, with 115 species (Table 4, Online Appendix 1), in contrast to Poaceae in Meter et al. (2002) (Table 6). However, this was consistent with the ranking of Fabaceae as the second most important species contributor to the KZN flora (Hilliard & Burtt 1987). There was one endemic (0.9%) and 11 near endemic species (9.6%) amongst the listed Fabaceae. Amongst the Fabaceae, the following species fell into ‘endangered’ categories in the RDL (Table 5): Crotalaria dura subsp. dura (near threatened) and Psoralea abbottii (vulnerable).

Using the old family structure Liliaceae for comparative purposes, this family is the third most species-rich family, with 109 species (Table 4). There were 5 endemic (4.6%) and 14 near endemic species (12.8%) amongst the listed Liliaceae. Amongst the Liliaceae, the following species fell into ‘endangered’ categories in the RDL (Table 5): Aloe kniphofioides (vulnerable), Aloe linearifolia (near threatened), Sandersonia aurantiaca (declining), Bowiea volubilis (vulnerable), Eucomis bicolor (near threatened), Eucomis comosa (declining) and Merwilla plumbea (near threatened).

The Orchidaceae was the fourth most species-rich family in 3029DA WEZA, with 89 species (Table 4, Online Appendix 1). There were three endemics and 25 near endemics amongst the listed Orchidaceae. Four orchid species collected were listed as rare. These were: Disa oreophila subsp. erecta (rare, endemic); Disa tysonii (rare, near endemic); Satyrium microrhynchum (rare, near endemic) and Schizochilus flexuosus (rare, near endemic).

The Iridaceae was the fifth most species-rich family in 3029DA WEZA, with 59 species (Table 4, Online Appendix 1) Three of these were endemic (5.1%) and 21 were near endemic (35.6%) (Table 5). One of these species was listed as rare, one as vulnerable, one as declining and one as endangered; they were: Dierama ambiguum (endangered, near endemic) (Table 5); Dierama tysonii (vulnerable, near endemic); Gladiolus oppositiflorus subsp. salmoneus (declining, near endemic) and Hesperantha ingeliensis (rare, near endemic).

The Poaceae ranked sixth, with 58 species in 3029DA WEZA (Table 4, Online Appendix 1). A comparison of the relative richness of Poaceae at Ngeli against that reported for Mahwaqa Mountain, the southern Drakensberg (Meter et al. 2002) and KZN as a whole suggest that Poaceae might have been under-collected at Ngeli (Table 6). Indeed, once a base species list was available, it proved to be relatively easy to collect previously uncollected grass species in the study area; there are undoubtedly further grass species awaiting collection there. This historical under-collecting of Poaceae could explain the difference in monocotyledon to dicotyledon ratio between the present study and that of Meter et al. (2002) (Table 6). One species was endemic (1.7%) and three were near endemic (5.2%) (Table 3). Poaceae included 5 naturalised species (Table 4, Online Appendix 1).There were no ‘endangered’ species in this list (Table 5, Online Appendix 1).

The Asclepiadaceae (57 species) (Table 4, Online Appendix 1), again used for purposes of comparison against older studies, was the seventh most numerous family in 3029DA WEZA. Three of these species were endemic (5.3%) and 17 near endemic (29.8%). Five species were listed as rare and three as vulnerable. These were: Asclepias xysmaloboides (rare, endemic); Aspidonepsis cognate (rare, near endemic); Brachystelma molaventi (vulnerable); Brachystelma tenellum (vulnerable); Emplectanthus gerrardii (rare); Schizoglossum elingue subsp. purpureum (rare, endemic); Schizoglossum montanum (rare, endemic); Schizoglossum singulare (vulnerable).¹

The Scrophulariaceae had 42 species in 3029DA WEZA (Table 4, Online Appendix 1). Of these, 4 were endemic (9.5%) and 11 were near endemic (26.1%) (Table 3). None of these Scrophulariaceae were listed as endangered in the RDL (Table 5).

The Euphorbiaceae in 3029DA WEZA had 32 species (Table 4, Online Appendix 1). Amongst these were two near endemic species (Table 5), one near threatened species (Alchornia hirtella forma glabrata) and one vulnerable species (Euphorbia flanaganii).

The Lamiaceae had 32 species in 3029DA WEZA (Table 4, Online Appendix 1). Amongst these were four near endemic species (Table 5). None of these Lamiaceae were in ‘endangered’ categories in the RDL, but there were two naturalised species (Table 5).

The 10 families above all comprised more than 2% of the species in the list (Online Appendix 1). Together they contributed 53.5% of the species and 41.9% of the genera.

Comparative data for this study, for Mahwaqa Mountain, for the southern Drakensberg (Meter et al. 2002) and for KZN as a whole, are shown in Table 6. As mentioned before, the Poaceae in 3029DA WEZA appeared to be comparatively under-represented, despite additional collecting. Fabaceae, in contrast, appeared to be over-represented when compared with Mahwaqa and the southern Drakensberg (Meter et al. 2002), although as a proportion of all recorded species, their data were more in line with the data for KZN as a whole. These apparent anomalies might justify a review of the Poaceae (some work already done) and Fabaceae to determine whether these relative proportions are real or as a result of collecting bias. The latter is unlikely given the significant number of collectors involved here.

The biodiversity and conservation value of the study area lie in the conserved pockets of eastern mistbelt forest and Drakensberg foothills moist grassland and eastern mistbelt grassland. This was evidenced by the number of species regarded as endangered (5), vulnerable (18), near threatened (10), critically rare (1), rare (20) or declining (11) amongst the 1554 species covered in the list, that is, more than 4% of the species are under some degree of threat. This, together with the fact that, amongst the angiosperms, there are 44 (3.1%) endemic species (11 monocotyledons and 33 dicotyledons) and 256 (17.8%) near endemic species (79 monocotyledons and 177 dicotyledons), reinforces the biodiversity and conservation value of the 3029DA WEZA area. Furthermore, this area supports important fauna (Adie et al. 2013; C. Potgieter [University of KwaZulu-Natal], pers. comm.). This necessitates that the responsible authorities ensure that the area is appropriately protected from invasive alien species, and uncontrolled and illegal use by the neighbouring communities (Adie et al. 2013). An interesting question to be investigated is whether all the species collected in the period leading up to 1891, when protection was afforded this and other KZN forests (McCracken 2004), are still present.